Github user mengxr commented on the pull request:
https://github.com/apache/spark/pull/2847#issuecomment-70027659
Had an offline discussion with @jackylk and here is the summary:
1. Keep only the parallel FP-Growth implementation, because it is generally
more efficient than Apriori, especially on medium/large datasets. @jackylk can
share some performance testing results.
2. Rename the package "fim" (frequent itemset mining) to "fpm" (frequent
pattern mining). There is no standard acronym this family of mining algorithms.
Frequent patten mining is a broader term than frequent itemset mining. This
package name is also used in Mahout.
3. Include links to the original FP-Growth paper and the PFP paper in the
doc.
4. Have FPGrowth take `minSupport` at a parameter and implement
`run(RDD...): FPGrowthModel`, where FPGrowthModel holds an RDD of frequent
itemsets and counts.
5. Hide methods used internally.
6. Update code style: a) remove extra empty lines; b) fix indentation; c)
change variable names; d) line width; etc.
7. Check whether we can use generic type for items (for Java API).
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
